During audio coding, considering the bit rate limitation and audibility characteristics of human ears, information of low-frequency audio signals is preferably coded and information of high-frequency audio signals is discarded. However, with the rapid development of the network technology, the network bandwidth limitation is being reduced. Meanwhile people's requirements for the timbre are higher and higher, and people desire to restore the information of the high-frequency audio signals by adding the bandwidth for the signals. In this way, the timbre of the audio signals is improved. In an embodiment, this may be implemented using bandwidth extension (BWE) technologies.
Bandwidth extension may extend the frequency scope of the audio signals and improve signal quality. At present, the commonly used BWT technologies include, for example, the time domain (TD) bandwidth extension algorithm in G.729.1, the spectral band replication (SBR) technology in moving picture experts group (MPEG), and the frequency domain (FD) bandwidth extension algorithm in International Telecommunication Union (ITU-I) G.722B/G.722.1D.
FIG. 1 and FIG. 2 are schematic diagrams of bandwidth extension. That is, no matter whether the low-frequency (for example, smaller than 6.4 kilohertz (kHz)) audio signals use TD coding or FD coding, the high-frequency (for example, 6.4-16/14 kHz) audio signals use TD-BWE or FD-BWE for bandwidth extension.
In the prior art, only TD coding of the TD-BWE or FD coding of the FD-BWE is used to code the high-frequency audio signal, without considering the coding manner of the low-frequency audio signal and the characteristics of the audio signal.